| Commit message (Collapse) | Author | Age | Files | Lines | ||
|---|---|---|---|---|---|---|
| ... | ||||||
| * | NamedMDNode is never used so there is no need to enumerate it here. | Devang Patel | 2010-01-09 | 1 | -1/+1 | |
| | | | | | llvm-svn: 93039 | |||||
| * | Fix nondeterministic behavior. | Julien Lerouge | 2010-01-09 | 1 | -6/+7 | |
| | | | | | llvm-svn: 93038 | |||||
| * | NamedMDNode element is either null or MDNode. Otherwise this is a malformed ↵ | Devang Patel | 2010-01-09 | 1 | -1/+3 | |
| | | | | | | | NamedMDNode record. llvm-svn: 93037 | |||||
| * | Derive NamedMDNode from Value. | Devang Patel | 2010-01-09 | 3 | -11/+25 | |
| | | | | | llvm-svn: 93032 | |||||
| * | Dan pointed out checking whether a node is dead by comparing its opcode to ↵ | Evan Cheng | 2010-01-09 | 1 | -8/+26 | |
| | | | | | | | ISD::DELETED_NODE is not safe. Use a DAGUpdateListener to remove dead nodes from work list instead. llvm-svn: 93031 | |||||
| * | Fix a critical bug in 64-bit atomic operation lowering for 32-bit. The ↵ | Evan Cheng | 2010-01-08 | 1 | -8/+10 | |
| | | | | | | | results of the cmpxchg8b instructions are being thrown away when it branches back to the top of the checking loop. This means the loop always compares against the old value and this can result in a dead lock. llvm-svn: 93028 | |||||
| * | Remove unnecessary dyn_cast and add a comment. Part of a WIP. | Eric Christopher | 2010-01-08 | 1 | -4/+4 | |
| | | | | | llvm-svn: 93026 | |||||
| * | mplement a theoretical fixme. | Chris Lattner | 2010-01-08 | 1 | -3/+7 | |
| | | | | | llvm-svn: 93024 | |||||
| * | rename CanEvaluateInDifferentType -> CanEvaluateTruncated and | Chris Lattner | 2010-01-08 | 1 | -71/+28 | |
| | | | | | | | simplify it now that it is only used for truncates. llvm-svn: 93021 | |||||
| * | Fix comment. | Evan Cheng | 2010-01-08 | 1 | -1/+1 | |
| | | | | | llvm-svn: 93020 | |||||
| * | teach instcombine to delete sign extending shift pairs (sra(shl X, C), C) when | Chris Lattner | 2010-01-08 | 1 | -4/+15 | |
| | | | | | | | the input is already sign extended. llvm-svn: 93019 | |||||
| * | fix PR5978 by peeling the loop so that we avoid shifting the | Chris Lattner | 2010-01-08 | 1 | -2/+2 | |
| | | | | | | | | result int by 8 for the first byte. While normally harmless, if the result is smaller than a byte, this shift is invalid. llvm-svn: 93018 | |||||
| * | Suppress an unused variable warning when assertions are off; | Duncan Sands | 2010-01-08 | 1 | -2/+3 | |
| | | | | | | | remove some trailing whitespace while there. llvm-svn: 93008 | |||||
| * | tidy up some stuff duncan pointed out. | Chris Lattner | 2010-01-08 | 1 | -2/+3 | |
| | | | | | llvm-svn: 93007 | |||||
| * | Minor change, change the order of two "let Inst{...}" stmts within multiclass | Johnny Chen | 2010-01-08 | 1 | -1/+1 | |
| | | | | | | | T2I_bin_ii12rs definition. llvm-svn: 93006 | |||||
| * | After further thought revert the patch to make fast-isel avoid | Eric Christopher | 2010-01-08 | 1 | -6/+0 | |
| | | | | | | | | | | | | putting relocations into the constant pool - this isn't needed for correctness and in the rare occasion it happens would pull us out of fast isel for the block. If fast-isel application startup time ever becomes an issue we can add better support for these addresses instead of bailing. llvm-svn: 92995 | |||||
| * | ReplaceAllUsesOfValueWith may delete other nodes that the one being ↵ | Evan Cheng | 2010-01-08 | 1 | -1/+2 | |
| | | | | | | | replaced. Do not delete dead nodes again. llvm-svn: 92988 | |||||
| * | Fix what looks to me obvious instruction definition bugs. | Evan Cheng | 2010-01-08 | 2 | -1/+3 | |
| | | | | | | | | 1. CMPXCHG8B and CMPXCHG16B did not specify implicit physical register defs and uses. 2. LCMPXCHG8B is loading 64 bit memory, not 32 bit. llvm-svn: 92985 | |||||
| * | Remove extraneous include. | Eric Christopher | 2010-01-08 | 1 | -1/+0 | |
| | | | | | llvm-svn: 92972 | |||||
| * | teach ComputeNumSignBits to look through PHI nodes. | Chris Lattner | 2010-01-07 | 2 | -7/+19 | |
| | | | | | llvm-svn: 92964 | |||||
| * | Enhance instcombine to reason more strongly about promoting computation | Chris Lattner | 2010-01-07 | 1 | -51/+144 | |
| | | | | | | | | that feeds into a zext, similar to the patch I did yesterday for sext. There is a lot of room for extension beyond this patch. llvm-svn: 92962 | |||||
| * | Fix rdar://7517201, a regression introduced by r92849. | Chris Lattner | 2010-01-07 | 1 | -1/+1 | |
| | | | | | | | | | | | When folding a and(any_ext(load)) both the any_ext and the load have to have only a single use. This removes the anyext-uses.ll testcase which started failing because it is unreduced and unclear what it is testing. llvm-svn: 92950 | |||||
| * | factor this code better and reduce nesting at the same | Chris Lattner | 2010-01-07 | 1 | -61/+41 | |
| | | | | | | | time, no functionality change. llvm-svn: 92948 | |||||
| * | APInt'fy TargetLowering::SimplifySetCC to fix PR5963. | Evan Cheng | 2010-01-07 | 1 | -15/+15 | |
| | | | | | llvm-svn: 92943 | |||||
| * | Simplify code. No intended functionality/performance change. | Benjamin Kramer | 2010-01-07 | 1 | -37/+12 | |
| | | | | | llvm-svn: 92938 | |||||
| * | If the data requires a relocation then don't attempt to | Eric Christopher | 2010-01-07 | 1 | -0/+7 | |
| | | | | | | | | add it to the constant pool for fast-isel. We already don't add it for the normal case. llvm-svn: 92934 | |||||
| * | We need to put any kind of data with a relocation into a | Eric Christopher | 2010-01-07 | 1 | -1/+1 | |
| | | | | | | | not-readonly segment on darwin. llvm-svn: 92933 | |||||
| * | Use separate namespace for named metadata. | Devang Patel | 2010-01-07 | 5 | -12/+40 | |
| | | | | | llvm-svn: 92931 | |||||
| * | constant materialization could be improved. | Chris Lattner | 2010-01-07 | 1 | -0/+33 | |
| | | | | | llvm-svn: 92921 | |||||
| * | Kill dead store. | Benjamin Kramer | 2010-01-07 | 1 | -2/+1 | |
| | | | | | llvm-svn: 92920 | |||||
| * | Remove dead variable. | Benjamin Kramer | 2010-01-07 | 1 | -7/+2 | |
| | | | | | llvm-svn: 92919 | |||||
| * | Use pop_back_val instead of back()+pop_back. | Benjamin Kramer | 2010-01-07 | 1 | -7/+4 | |
| | | | | | llvm-svn: 92918 | |||||
| * | Use a do-while loop instead of while + boolean. | Benjamin Kramer | 2010-01-07 | 1 | -6/+4 | |
| | | | | | llvm-svn: 92912 | |||||
| * | Be less stingy as to how many selects and phi nodes we | Duncan Sands | 2010-01-07 | 1 | -2/+2 | |
| | | | | | | | are prepared to look through. llvm-svn: 92898 | |||||
| * | More trivial optimizations to a function well outside the critical path | Douglas Gregor | 2010-01-07 | 1 | -15/+8 | |
| | | | | | llvm-svn: 92896 | |||||
| * | handle ConstantVector while I'm in here. | Chris Lattner | 2010-01-07 | 1 | -0/+3 | |
| | | | | | llvm-svn: 92892 | |||||
| * | fix a globalopt crash on 'bullet' (handling evaluation of a store | Chris Lattner | 2010-01-07 | 1 | -16/+20 | |
| | | | | | | | | | | | to an element of a vector in a static ctor) which occurs with an unrelated patch I'm testing. Annoyingly, EvaluateStoreInto basically does exactly the same stuff as InsertElement constant folding, but it now handles vectors, and you can't insertelement into a vector. It would be 'really nice' if GEP into a vector were not legal. llvm-svn: 92889 | |||||
| * | Fix a minor regression from my dag combiner changes. One more place which ↵ | Evan Cheng | 2010-01-07 | 1 | -4/+12 | |
| | | | | | | | needs to look pass truncates. llvm-svn: 92885 | |||||
| * | Switch StringRef::edit_distance over to using raw pointers, since both | Douglas Gregor | 2010-01-07 | 1 | -7/+25 | |
| | | | | | | | | std::vector and llvm::SmallVector have annoying performance tradeoffs. No, I don't expect this to matter, and now it won't. llvm-svn: 92884 | |||||
| * | 80 column violations | Jim Grosbach | 2010-01-06 | 1 | -4/+6 | |
| | | | | | llvm-svn: 92876 | |||||
| * | Allow double defs after tail duplication. | Jakob Stoklund Olesen | 2010-01-06 | 1 | -1/+2 | |
| | | | | | llvm-svn: 92874 | |||||
| * | Add Target hook to duplicate machine instructions. | Jakob Stoklund Olesen | 2010-01-06 | 4 | -22/+56 | |
| | | | | | | | | Some instructions refer to unique labels, and so cannot be trivially cloned with CloneMachineInstr. llvm-svn: 92873 | |||||
| * | Addressing mode 6 (load/store) instructions can't encode an immediate offset | Jim Grosbach | 2010-01-06 | 1 | -1/+4 | |
| | | | | | | | for stack references. llvm-svn: 92871 | |||||
| * | Anti-dependency breaking needs to be careful regarding instructions with | Jim Grosbach | 2010-01-06 | 2 | -3/+9 | |
| | | | | | | | multiple register definitions. llvm-svn: 92864 | |||||
| * | Move the object size intrinsic optimization to inst-combine and make | Eric Christopher | 2010-01-06 | 2 | -24/+12 | |
| | | | | | | | it work for any integer size return type. llvm-svn: 92853 | |||||
| * | Comment. | Evan Cheng | 2010-01-06 | 1 | -0/+3 | |
| | | | | | llvm-svn: 92850 | |||||
| * | Teach dag combine to fold the following transformation more aggressively: | Evan Cheng | 2010-01-06 | 5 | -114/+295 | |
| | | | | | | | | | | | (OP (trunc x), (trunc y)) -> (trunc (OP x, y)) Unfortunately this simple change causes dag combine to infinite looping. The problem is the shrink demanded ops optimization tend to canonicalize expressions in the opposite manner. That is badness. This patch disable those optimizations in dag combine but instead it is done as a late pass in sdisel. This also exposes some deficiencies in dag combine and x86 setcc / brcond lowering. Teach them to look pass ISD::TRUNCATE in various places. llvm-svn: 92849 | |||||
| * | When parsing function-local metadata, create a function-local MDNode | Victor Hernandez | 2010-01-06 | 2 | -10/+16 | |
| | | | | | llvm-svn: 92838 | |||||
| * | 80 column and whitespace cleanup | Jim Grosbach | 2010-01-06 | 4 | -95/+108 | |
| | | | | | llvm-svn: 92837 | |||||
| * | Fix a README item: have functionattrs look through selects and | Duncan Sands | 2010-01-06 | 2 | -40/+42 | |
| | | | | | | | | | | phi nodes when deciding which pointers point to local memory. I actually checked long ago how useful this is, and it isn't very: it hardly ever fires in the testsuite, but since Chris wants it here it is! llvm-svn: 92836 | |||||

